Cloud Analytics is a dynamical framework for high performance big data analytics. Together with high performance statistical applications it is intended to bring the capabilities of specialized high performance computing (HPC) hardware to SMEs, providing invaluable information extraction services at very low cost by means of highly parallelizable algorithms running on our private cloud.
SVMs analyse data and recognize patterns, with linear and kernel classification and regression analysis. Given a set of training samples, Cloud Analytics will create a predictive model following the given parameters. The obtained modelled function can then be used independently to predictively classify any future data instantly.
Random Forest is an inherently parallelizable method for categorization that Cloud Analytics makes use for classifying structured data. The resulting model is very compact, described in only a handful of conditionals, and can be used for very fast classification.
Cloud Analytics makes use of modified K-means, C-means and K-medoids for fast parallel clustering of data of which we know nothing about. The resulting clusters can tell us a lot about the nature of the data. The model of resulting centroids can be used to classify additional data.